Dev/optimize merkle root

[Jeanmichel7]

Test	Before	After
full_169	134001057	113894412
full_757738	9372119456	6850847496

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
raito	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Sep 26, 2024 11:54pm

[maciejka]

Why you don't use Box<Digest> instead of Box<[u32;8]>?

[Jeanmichel7]

Why you don't use Box<Digest> instead of Box<[u32;8]>?

Because the hash return is Box::<[u32,8]>,
this avoids unnecessary conversions such as
next_hashes.append(BoxTrait::new(DigestTrait::new(hash.unbox())));

I've tried it, and the performance loss is very small 6851088696 instead of 6850847496
validation>block.cairo is a little more readable,
utils>merkle_tree.cairo is a little less readable

[m-kus]

We cannot use sha256 syscall from Starknet

[Jeanmichel7]

That's right..., I'll have to see how I can use the local sha256.